OK, I am going to check at home...

Hi Hao

I did get your email in the evening but I didn't get your def file...So your turb. modell was SST. What were the boundary conditions? The velocity, ref. pressure etc. Is the Turbulence transition active in your simulation? Do you use stuctured hexa mesh?

Sorry, Roland. It seems that I sent the email before the file was finished uploading. So, I sent a new email and got the def file attached. Please check it and give me some advice. Thanks a lot. By the way, in the file I sent to you, I did not activate transition model. But I really did it in the other simulations later. It turns out the residue began to oscilate after a period of time and did not converge any more (~1e-3). Some guy suggested a transient simulation.

I experienced the periodic oscillation in my calculation too. I think that this phenomena is absolutely normal. The boundary layer separates around 14 degree. This is a short transient process, which is about 0.4-0.6 sec in time. It is evident that the solver can not calculate this short transient period because the simulation type is steady state therefore the calculation can not converge. If you increase the stall angle a little than a steady separation is going to develop behind the wing so the periodic oscillation is going to cease in the convergence history. If you do a transient simulation for this critical stall angle so you are going to see the process of the separation. (It is worth trying because this is a very interest and spectacular phenomena.) But you are going to see the critical angle-position clearly even if you calculate a simple steady state process, you must plot the normal forces on the wing in the solver manager during the running...

Hello Hao,

I studied the y plus at 0 degree in your file, and the average was about 5-6 at the wing. You must divide the first cell along the wing. In addition: The turbulence kinetic energy distribution is next to wing high. This is the effect of the boundary layer. The first 4-5 cells must contain this large turb. kinetic energy gradient range anyway. So the most important conditions are: y+ < 1 and correct dividing of the range of the turb.kinetic energy increase. But there is a thing which is very important: What was for you the lift force and the drag force during the evaluation? F_y = Lift force and F_x = Drag force?

NO, the lift force is calculated by F_y*cosAOA - F_x*sinAOA. The lift force is perpendicular to the free stream direction, right? Roland, what did you use for aspect ratio? less than 1.1?

Hello, Roland. Would you mind sending me one copy of your "lift (and drag) coefficient vs angle of attack" at RE~360000? I took your advice and got the stall angle around 14 degree. However, when i ran the steady-state simulation at AOA=15, the case did not converge (but oscilate). Did you get those CL and CD in steady-state simulations? Cheers. Hao

[Zmur]

Originally Posted by Attesz
;92143

an example:

http://kepfeltoltes.hu/view/090303/i...toltes.hu_.jpg

attesz

Hello.
How do you create such meshes? Is it in CFX-mesh?

[Roland R]

Hello Hao

I sent a mail to your e-mail adress last week. Did you get it?

Roland

[cfdmaster]

could someone create a cfx mesh for naca 0009 with a flat bottom. I have made it in catia but i get a weird mesh using icemcfd

any help plz.

here is the airfoil

http://rapidshare.com/files/211377803/0009.igs.html